Project Experience

40+ Years. Anonymous. Concrete.

No glossy references, no inflated case studies. Real projects, real problems, real solutions – anonymised out of discretion, unabridged in substance.

Historical projects are presented anonymously. Client names, industries and timeframes are deliberately kept open – anyone needing specific details is welcome to raise it in conversation.
✈️
Delphi Aviation Real-time

Major European aviation group

Cockpit Software for Take-off Performance Optimisation

The Java development team was unable to deliver the required software on time. The task was taken over at short notice and implemented in Delphi – on schedule, deployment-ready and certification-capable. The software calculated weight-dependent performance parameters directly in the cockpit.

Sometimes it is not the technology that decides – it is who delivers. This project shaped my conviction: the right tool in experienced hands beats the "modern" tool in overwhelmed ones.

Delphi Sports Long-term

International golf software solution

Industry-Specific Management Software – 6 Years

Nearly six years contributing to a specialised software solution for the golf industry. Deep domain knowledge, mature architecture, close collaboration with subject matter experts. A project that demonstrates: software quality comes from continuity – not fast sprints.

Six years in one domain means: you understand the client's language. That is the foundation for software that truly fits.

🍽️
iOS / iPad Cross-platform B2B

Restaurant chain with multi-site structure

Inventory Management & B2B iPad App for the Hospitality Sector

Custom inventory management with integrated reporting, plus a cross-platform B2B iPad application for direct use by restaurateurs. End users were enthusiastic – the project was discontinued regardless.

Not every good piece of software survives internal decision-making. This experience sharpened my eye for the non-technical risks of a project – and for how important stakeholder management really is.

🔧
Delphi Excel Integration Data Security

Representative of a globally leading engine manufacturer

Cost Planning with Highly Sensitive Client Data

Complex integration between a Delphi application and extensive Excel structures for customer cost planning in the aircraft engine sector. The greatest requirement was not the technical complexity – it was the absolute confidentiality of live data from competitors.

Sensitive data does not belong in the cloud – that was true then for Excel files, and it is true today for AI models. Data sovereignty is not a new topic. Only the technology has changed.

💳
Delphi Fintech Consulting

Leading provider of card-based loyalty systems

Architecture Consulting under Banking Security Standards

Consulting mandate for a Delphi application at a company operating in a regulated payments environment – with security requirements at banking level. Access, processes and infrastructure were correspondingly restrictive.

Being able to work in highly regulated environments means: understanding compliance, exercising discretion, and the ability to advise effectively despite constraints.

🏛️
Cross-platform Document Management B2B

Architecture firm with client project management

Digital Plan Archive for Construction Projects

Cross-platform application for the structured storage and management of plans, images and project documents – organised and accessible per client project. The project was discontinued because fundamental architectural decisions could not reach consensus.

Software architecture is not an opinion – it is a technical matter. Where core convictions diverge too far, quality suffers in the long run. This experience reinforced my insistence on clear architectural principles from the outset.

Control Software Motion Control Industrial

Manufacturer of high-performance motion solutions

Control Software for High-Performance Drive Systems

Development of control software for high-precision motion control solutions in the industrial sector. The engagement concluded as planned when the underlying hardware architecture underwent a complete redesign – the project moved in a new technical direction internally.

Control software is not a standard product – it must map the hardware reality exactly. When the hardware changes fundamentally, the software changes from the ground up too. That is not failure – it is engineering reality.

What we build today

Alongside consulting, AINA develops its own tools and services – in active use, openly documented, and living proof of what AI-assisted software engineering can achieve today.

🔄
Rust ratatui (TUI) Tauri (GUI) macOS

AINA Technology – Own project

app-updater

A modular update system for macOS applications built in Rust – with a terminal UI via ratatui for the command line and a native GUI via Tauri for desktop use. From concept to production-ready tool in one week, developed using AI partner coding. Plugin architecture enables unlimited extensibility without touching the core.

A concrete example of what AI partner coding delivers: clean architecture, a real result, no prototype – in a fraction of the usual development time.

View Showcase →
🌐
HTML / CSS Accessibility SEO Multi-page

AINA Technology – This website

aina.technology

Fully self-conceived, architecturally considered and realised using AI partner coding: a static, accessible multi-page system without CMS overhead. Accessibility-compliant design, structured sitemap, integrated blog system and a consistent separation of content, structure and style – all in-house, legally sound and maintainable.

Anyone offering software architecture and AI-assisted development as a service should be able to demonstrate both in their own projects – not just describe them.

🧠
MLX Apple Silicon LLM Fine-Tuning

AINA Technology – Evaluation infrastructure

Local LLM Infrastructure on Apple Silicon

Building and operating a proprietary evaluation infrastructure for local language models on Apple Silicon via MLX – from model selection through performance benchmarking to practical quality assessments for real-world use cases. Findings and insights feed actively into the open-source community.

Anyone recommending local AI solutions must have operated them, evaluated them and pushed them to their limits. This infrastructure is the foundation for that.

Community contribution on GitHub →
⚙️
MCP AI Infrastructure Context Management

AINA Technology – In active use

PixelMemory & Remote MCP Gateway

Proprietary tools for persistent AI context management and the secure remote integration of MCP servers – born from our own daily needs in AI-assisted work. Both systems are in active production use and continuously evolving.

Anyone advising on AI infrastructure should have built and battle-tested it in everyday use themselves. PixelMemory is not a demo – it is the tool AINA works with every day.

🎯
Fine-Tuning Local Models Cloud GPU Data Sovereignty

AINA Technology – Service

Model Fine-Tuning for Specialised Use Cases

Fine-tuning language models on your own data and domain-specific requirements – locally on dedicated hardware for maximum data sovereignty, or scalably via cloud GPU infrastructure. Not an off-the-shelf model, but a model that masters the language, tone and expertise of your organisation.

A generic model answers generically. A model tailored to your organisation answers like an experienced employee – precise, consistent, with the right context.

Ready to get specific?

In a personal conversation, I am happy to go into individual projects in more depth.

Request a conversation →